CT Dose calculations for individual patients – what you

should know

Elly Castellano

2CTUG 2010

Does this put you off?

3CTUG 2010

Objective

• how accurately can ImPACT CT dose calculator and CT-Expo calculate individual patient effective dose (ICRP 60)?

4CTUG 2010

The patient, the code and the error

• modified Zubal voxelised adult phantom– breasts and ovaries added– all radiosensitive organs included– variable dimensions

5CTUG 2010

The patient, the code and the error

• RMH-Linköping Monte Carlo model– SSCT and MSCT scanners– geometry, spectrum, beam

shaping filter, couch (OFF)– axial or helical scanning

• validated against experiment– < 10 %

zx

y

FID

voxel phantom

x-ray beam

ρ

αβ

beam shaping filter

zx

y

FID

voxel phantom

x-ray beam

ρ

αβ

beam shaping filter

zx

y

FID

voxel phantom

x-ray beam

ρ

αβ

beam shaping filter

6CTUG 2010

How does RMH-Link compare against NRPB and GSF?

• two scanner models: HiSpeed CT/i SSCT and LightSpeed 16 MSCT• whole body irradiation• compared against ImPACT calculator and scaled CT-Expo

• differences due to organ modelling• similar results for LightSpeed 16• effective doses agree to within 7 %

-40

-30

-20

-10

0

10

20

30

40

50

tes

tes

lun

g

sto

ma

ch

blad

de

r

bre

as

t

live

r

organ

% d

iffer

ence

fro

m R

MH

-Lin

k re

sults Cristy

Adam

Eva

-40

-30

-20

-10

0

10

20

30

40

50

ova

ries

rbm

colo

n

oes

oph

agu

s

thyr

oid

organ

% d

iffer

ence

fro

m R

MH

-Lin

k re

sults Cristy

Adam

Eva

HiSpeed CT/i

7CTUG 2010

How do I set the scan range?• four strategies

– anatomical landmarks– scan range– fractions of irradiated

organs– NRPB technique

• which one?

• simulate– brain, thorax, abdomen,

pelvis scans– one scanner: HiSpeed CT/i

(a)

(b)

8CTUG 2010

How do I set the scan range?

• effective doses for male phantom– similar results for female and hermaphrodite phantom

• matching fraction of irradiated organs most accurate– agreement within 20 %

-50-40-30-20-10

01020304050

brain thorax abdomen pelvis

CT scan

% d

iffer

ence

fro

m R

MH

-L

ink

(mS

v)

anatomy

range

organ fraction

-50-40-30-20-10

01020304050

brain thorax abdomen pelvis

CT scan

% d

iffer

ence

fro

m R

MH

-L

ink

(mS

v) anatomy

range

organ fraction

NRPB

scaled CT Expo ImPACT dose calculator

9CTUG 2010

How do I set the scan range?scan landmarks start cm end cm

brain base of skull to vertex

81 93

thorax lung apices to bases

36 68

abdomen dome of diaphragm to iliac crest

17 47

pelvis iliac crest to symphysis pubis

-1 17

10CTUG 2010

Do I need to allow for helical over-ranging?

• measure over-range• from total exposure time and scan

parameters

• add over-range to scan range

• simulate– one scanner: LightSpeed 16– brain, thorax, abdomen, pelvis

scans– helical mode with varying pitch– axial mode without over-ranging

• equivalent to dose calculators

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

0.5 0.75 1 1.25 1.5 1.75 2

pitch

irra

dia

ted

- im

aged

ran

ge

mm

7.5 mm

5 mm

3.75 mm

2.5 mm

1.25 mm

LightSpeed 16

11CTUG 2010

Do I need to allow for helical over-ranging?

• helical v axial comparison– 4 to 13 % discrepancy – depends on

radiosensitivity of boundary organs 0

1

2

3

4

5

6

7

8

0 1 2 3 4 5 6 7 8

E (mSv) helical simulation

E (m

Sv)

axi

al c

alcu

latio

n

brain

thorax

abdomen

pelvis

pitch

RMH-Link model only

12CTUG 2010

Do I need to allow for helical over-ranging?

• comparison with MC calculators– agreement better than

20 % for thorax, abdomen and pelvis scans 0

2

4

6

8

10

12

0 2 4 6 8 10 12

E using RMH-Link CT model (mSv)

E u

sin

g c

om

me

rcia

l pac

kag

e (m

Sv)

Cristy

Cristy male

Cristy female

Adam

Eva

pitch

abdomen scan

13CTUG 2010

Do I need to adjust for patient size?

• simulate– three phantoms

• 50, 70 and 90 kg

– one scanner• HiSpeed CT/i

– brain, thorax, abdomen, pelvis scans– axial scanning– fixed exposure parameters

14CTUG 2010

Do I need to adjust for patient size?

• effective dose increases in smaller patients

• 13 % change in effective dose for 30 % change in weight

scan effective dose ratio

50 kg 70 kg 90 kg

brain* 0.99* 1.00 0.94

thorax 1.13 1.00 0.87

abdomen 1.10 1.00 0.87

pelvis 1.13 1.00 0.91

* scaled separately

15CTUG 2010

And what about mA modulation?

• retrospective patient dose survey

• 30 patients

• TAP protocol– 120 kV, 20 mm collimation, 1.375 pitch, 7.5/7.5

mm slices

• data collection– mA, scan range from images– DLP from patient log book

16CTUG 2010

And what about mA modulation?

• individualised effective dose calculation

• RMH-Link MC model– resize voxel phantom to each patient– assume scan range set on landmarks consistently

• include helical over-ranging

– calculate conversion factor per rotation– estimate average mAs per rotation

• from images

– scale conversion factor by mAs per rotation – sum contribution from all rotations

17CTUG 2010

And what about mA modulation?

• best approach with the ImPACT dose calculator– Cristy phantom divided into anatomical regions:

• 7 regions: shoulders, lungs, lung / liver overlap, liver, bowel, pelvis, femora

• 3 regions: thorax, abdomen, pelvis• 1 region: torso

• conversion factors calculated for each region• average mAs estimated for each region

– from images

• average mAs estimated from DLP• effective dose corrected for patient size

18CTUG 2010

And what about mA modulation?

• ImPACT dose calculator overestimates effective dose by 13 – 19 % on average– scanner matching is a factor

• average mAs for the scan provides sufficient accuracy

0

5

10

15

20

25

30

5 10 15 20 25

effective dose using RMH-Link CT model (mSv)

effe

ctiv

e d

ose

usi

ng

ImP

AC

T te

chn

ique

(mS

v)

7 regions

3 regions

+/- 25%

0

5

10

15

20

25

30

5 10 15 20 25

effective dose using RMH-Link CT model (mSv)

effe

ctiv

e d

ose

usi

ng

ImP

AC

T

tech

niq

ue

(mS

v)

1 region

DLP

+/- 25%

19CTUG 2010

What you should knowsource of error error optimisation technique optimised

error

Monte Carlo codes 7 % none 7 %

scan range 40 % match fraction of irradiated organs

20 %

helical over-ranging 13 % for 16-SCT add helical over-range ~ 0 %

patient size 13 % for 30 kg deviation

correct for patient size ~ 0 %

mA modulation 13 %* average mAs or CTDIvol 13 %*

combined 46 % 25 %

* comparison of Link-RMH effective doses with and without tube current modulation

20CTUG 2010

Conclusions• effective doses can be calculated with 25 %

accuracy if– set scan range by matching fractions of organs

irradiated– include helical over-range– correct for patient size– use average mAs for the scan

• effective doses can be calculated with 45 % accuracy for 50 – 90 kg patients otherwise